First of all, the answer lies in the very principle of a measurement. When you measure something (anything, even weight, length etc.) you actually compare the measured parameter of the object to another object (a standard) with known parameters. The same applies to radioactivity (cps) or any other parameter/measurement for that matter. This also means that it's impossible to take an "absolute" measurement of anything.
The other problems you seem to neglect is the problem of noise and aging/change of your measuring instrument.
Noise is something that always has to be accounted for, otherwise you'll end up "measuring" values which bear no meaning whatsoever (as they contain random "information" that's the noise itself) or end up with "glitches" in your measurements (and might even draw wrong conclusions from them). To translate this into your question: one count might not always be one count, because of the noise present in the instrument (both in the physical and the electronic part).
Aging is something that's particularly of concern with equipment dealing with radioactive radiation. Since such radiation is an "ionizing" one, it means that the very exposure to such source of radiation causes electrons to be kicked out of the instrument's material (which sometimes leads to permanent changes in the material structure). The other cause of aging of these materials is the very environment we live in. Just consider this: ~21% of the atmosphere we live in consists of a very reactive gas called oxygen. The other problem lies in the fact that literally everything is covered in a thin layer of water (a layer so thin that it can't be seen with the naked eye but causes problems in atomic force microscopy). Since most of the elements on Earth are at least somewhat reactive (with the exception of gold and platinum), they enter into chemical reactions with the water, the oxygen or other materials in the air (CO2, the sulfides causing the tarnishing of silver etc.) and thus the materials change as the result of these reactions. This will cause instruments to "go out of tune", their signal-to-noise ratio to decrease and other effects that cause instruments to go out of calibration.
Moreover Geiger counters are really precise instruments that are some of the most sensitive devices out there. This means that even really small changes will have a noticeable effect on their calibration.